Modeling cover crop biomass production and related emissions to improve farm-scale decision-support tools

CONTEXT: Cover crops deliver numerous ecosystem services including the capacity to sequester and store atmospheric CO2 offering promise as a natural climate solution. Farm system models must be able to accurately represent cover crop systems and estimate the associated net greenhouse gas emissions from this practice across agricultural contexts. OBJECTIVE: The objectives of this study were to: (1) parameterize new and existing cover crop species for DayCent - the biogeochemical model behind the decisions-support tool, COMET-Farm - using published data on cover crop biomass production, (2) validate differences between cover crop and control systems using published and long-term data from across the U.S. for SOC and N2O emissions, and (3) evaluate model performance and discuss improvements to cover crop data reporting to facilitate future testing. METHODS: We used published field experiment datasets from across the U.S. to estimate cover crop biomass production, SOC, and soil N2O emissions. We parameterized one new cover crop species in the model - sunn hemp (Crotalaria juncea L.) - and revised the parameters for existing model cover crop species using new biomass observations. We also validated the ability of the model to estimate changes in SOC and soil N2O emissions relative to a no cover crop control using empirical observations. RESULTS AND CONCLUSIONS: For data sets used in parameterization, DayCent reasonably captured observed trends in aboveground biomass C, N, and C:N ratios across six different cover crop species and mixtures, explaining 25-69%, 14-58%, and 29-50% of the variation, respectively, with a few exceptions. Using a smaller validation dataset for a subset of these cover crop species - cereal rye (Secale cereale L.) and vetch (Vicia spp.) - the model demonstrated less agreement between modeled and observed aboveground biomass. For SOC model validation, the observed mean SOC difference between cover crop and control treatments was 178.4 g m(-2) and the mean modeled difference was 158.2 g m(-2). Observed soil N2O emissions were highly variable impacting model performance, but DayCent did capture the general trend from field measurements for cover crops to suppress soil N2O emissions. SIGNIFICANCE: Our results demonstrate that even under limited data availability, process-based models like DayCent can inform on farm-scale biogeochemical processes under cover cropping. These findings are critical as cover crops are more widely adopted in the United States for their climate and environmental benefits. However, more detailed reporting from empirical studies will improve model estimates and reduce uncertainty, particularly data pertaining to farm management, soil, climate, and location.

Automated summary

This study aimed to parameterize new and existing cover crop species in the DayCent model, validate differences between cover crop and control systems using published data, and evaluate model performance and improvements in cover crop data reporting. The results showed that process-based models like DayCent can inform on farm-scale biogeochemical processes under cover cropping, but more detailed reporting from empirical studies is needed to improve model estimates and reduce uncertainty.